Introduction

show larger differences in magnitudes of depletion, which we interpret as changes in the remote effects onδD^precip during the different climate states.

4.2 Introduction

Tropical African climate is governed by the seasonal migration of the Intertropical Convergence Zone (ITCZ), a band of deep convection at the location of maximum insolation and moisture convergence. This causes northern and southern belts of monsoonal climates with rains in summer. The rainfall distribution over Africa de-pends on shifts in the large scale wind fields at different levels in the troposphere, sea-surface temperature (SST) variations (Giannini et al.,2003), atmospheric aerosols and changes in land surface boundary conditions (Nicholson,2000a;Zhao et al.,2011).

Equatorial Africa has a humid climate with two rainfall maxima per year. The north-ern and southnorth-ern temperate regions of the continent are affected by the equatorward displacement of the mid-latitude westerlies during the respective winter season. The subtropical deserts (the Sahara in the Northern Hemisphere and the Namib coastal desert in South West Africa) are dominated by subtropical anticyclones throughout the year.

The seasonal rainfall over the Sahel is strongly influenced by variations of SST anoma-lies in the tropical Atlantic, central and east Pacific and the Indian Ocean (Lamb, 1978;Palmer,1986;Hastenrath,1990;Bader and Latif,2003). Nicholson(2001) finds that the wet years in the Sahel are preceded by an anomalously warm tropical At-lantic. A model study byRowell(2003) finds that Sahel rainfall is also related to SST anomalies in the Mediterranean Sea. The African easterly jet (AEJ) is an important wind structure over northern Africa prominent in summer and helps to generate and maintain the wave disturbances that modulate the rainfall field in North West Africa (Cook, 1999; Thorncroft and Blackburn, 1999; Nicholson, 2001). Since the AEJ is

4.2. Introduction

associated with the divergence of moisture below the level of condensation, a strong jet is assumed to be connected with low precipitation over the western Sahel (Cook, 1999;Nicholson, 2001).

A prominent part of austral summer (December-February) rainfall over much of South Africa (SA) is derived from tropical-temperate trough systems that extend over continental SA and the adjacent SW Indian Ocean (Washington and Todd,1999;Todd et al.,2004). A second convergence zone, the Congo Air Boundary (CAB), separates the on-land flow from the Atlantic and Indian Oceans, thereby dividing easterly trades and westerly monsoonal wind systems over Africa (Nicholson,2000b).

Precipitation in eastern Africa is controlled by the position of two convergence zones, the ITCZ and the CAB and topography (Nicholson,1996). Ummenhofer et al.

(2009) find that positive SST anomalies in the western Indian ocean cause enhanced atmospheric moisture content. Consequently the westerly airflow gets strengthened over Central Africa and a strong anomalous convergence of moisture occurs over most of equatorial East Africa. Modeling studies (Latif et al.,1999;Friederichs and Paeth, 2006;Bader and Latif, 2011) also emphasize the relation between Indian ocean SST anomalies and precipitation over East Africa.

Three major time periods during the late Quaternary, i.e., the mid-Holocene (6 ka BP), Last Glacial Maximum (LGM, approximately 23-19 ka BP) and Heinrich stadial-1 (HSstadial-1, approximately stadial-18-stadial-16 ka BP), present a fitting scenario to study the impact of changes in the atmospheric and oceanic boundary conditions on the continental climate and hydrological cycle over Africa in the past. Climate modeling studies suggest a generally drier climate over tropical Africa during LGM and HS1 (Braconnot et al., 2000; Mulitza et al., 2008) and a wetter climate during the mid-Holocene (Joussaume et al., 1999; Doherty et al., 2000; Texier and Noblet, 2000; Braconnot et al., 2000;Zhao et al.,2005) than the present day climate.

During the early to mid-Holocene, tropical Africa was wetter than today (Gasse,

4.2. Introduction

2000), which is explained by the intensification and northward expansion of the African-Asian monsoon (Braconnot et al.,2007). A high Northern Hemisphere (NH) summer insolation in the early and mid-Holocene enhanced the thermal contrast be-tween land and sea producing stronger summer monsoons (Prell and Kutzbach,1987;

Hewitt and Mitchell,1996;deMenocal et al.,2000) as the continental surfaces respond faster to the seasonal cycle of insolation because of the low thermal capacity of the land. A multi-model analysis of the role of the ocean in the African monsoon during the mid-Holocene (Zhao et al., 2005) finds that the dipole (higher temperature to the north of 5^◦N and lower temperature to the south) in late summer SST anoma-lies in the tropical Atlantic increases the length of the African summer monsoon and precipitation over the Sahel. The Paleo Climate Modeling Intercomparison Project (PMIP) (Braconnot et al., 2000) finds that the ITCZ in the north only experienced small northward or southward shifts between the LGM and mid-Holocene periods.

The modeling study by Kutzbach et al. (1996) asserts the need to include not only orbital forcing but also the vegetation feedback in modeling the mid-Holocene cli-mate. During PMIP2, this has been addressed for the first time using comprehensive climate models (Braconnot et al., 2007).

Heinrich Events were six millennial-timescale, abrupt cool episodes during the late Pleistocene around the North Atlantic as a result of massive ice-berg discharge to the ocean (Heinrich, 1988; Broecker et al., 1992; Hemming, 2004). The Heinrich Events, which are identified as H1 through H6 from the youngest to the oldest, were followed by strong SST and salinity reductions in the North Atlantic (Bond et al., 1992; Vidal et al., 1997), which is assumed to have caused the slow-down of the Atlantic meridional overturning circulation (McManus et al.,2004;Brady and Otto-Bliesner, 2011). Climate models and proxies also show that the monsoon circulation was altered during the Heinrich events (Pausata et al.,2011b;Stager et al.,2011).

During the LGM, paleo-proxy records indicate generally dry conditions in both

4.2. Introduction

hemispheres (Shi et al., 1998; Prentice and Jolly, 2000; Wu et al., 2007), while from orbital forcing increased summer precipitation in the southern tropics is predicted.

Reduced SSTs in the tropics (MARGO, 2009) might have played a significant role in the tropical climate by reducing the evaporative flux (Gasse(2000) and references therein). Proxy studies suggest a dry central equatorial Africa during the LGM (Sche-fuß et al., 2005; Tierney et al., 2008). A study by Schefuß et al. (2005) using plant wax hydrogen isotope data and an alkenone-based SST reconstruction finds that pre-cipitation in Central Africa during the past 20,000 years was mainly controlled by the difference in sea surface temperatures between the tropics and subtropics of the South Atlantic Ocean, which opposes the assumption that the moisture availability in Central Africa was determined by the position of the ITCZ alone. In southeast-ern Africa, a climate reconstruction by Schefuß et al. (2011) suggests that remote atmospheric forcing by cold events in the northern high latitudes is a major driver of hydro-climatology, in contrast to the Indian ocean SST variability proposed byStager et al.(2011).

Stable isotopes of water in precipitation can be used to track the hydrological cycle because of the dependance of water isotopes on equilibrium and kinetic fractionation associated with the phase transitions of water. The isotope ratio is represented asδ value in with δ=(RSAMPLE/RVSMOW-1)× 1000, where RSAMPLE is the ratio of the heavier isotope to the lighter isotope of the sample and RVSMOW=155.76×10⁻³ is the hydrogen isotopic ratio of Vienna Standard Mean Ocean Water. The fractiona-tion of the stable isotopes of hydrogen and oxygen is strongly influenced by climate (Dansgaard, 1964). Rainfall amount, moisture source, altitude, distance from the coast, and humidity are the major factors influencing the stable isotope ratios of wa-ter in precipitation in the tropics (Dansgaard,1964;Rozanski et al.,1993;Gonfiantini et al., 2001).

In the monsoon domains, the ratio of isotopes in precipitation is mostly related to

4.2. Introduction

the amount of precipitation (amount effect; Dansgaard, 1964; Rozanski et al., 1993;

Dettman et al., 2001; Lee and Fung, 2008; Risi et al., 2008b). Vuille et al. (2003) suggest that the amount effect is caused by the small-scale vertical convection asso-ciated with precipitation in the tropics. As the condensation proceeds, the heavier isotopes are preferentially removed and the relative abundance of the heavier iso-topes in the water vapor decreases. Since the total amount of precipitation increases with the increase in the convective nature of a particular rainfall event, the isotopic composition of the precipitation gets more depleted. This effect can be amplified by the fact that isotopic exchange with water vapor and evaporative enrichment of raindrops are both greatly reduced with heavy rains (Dansgaard, 1964). Contrary to Vuille et al. (2003), Risi et al. (2008a) conclude that the predominant processes leading to the amount effect in the tropics are related to the fall and reevaporation of the precipitation, rather than processes occurring during the ascent of air parcels.

They further find that the fractionation process during rain fall contributes the most to the amount effect in regimes of weak precipitation, and the injection of vapor from the unsaturated downdraft is predominant in regimes of strong precipitation.

Another factor that affects the isotope ratios over land is continentality, also re-ferred to as the distance-from-coast effect. As oceanic air masses move inland and lose water through precipitation, the remaining atmospheric water vapor becomes progressively depleted in heavy isotopes and when the air mass reaches an orographic obstacle, the altitude effect (Gonfiantini et al., 2001) will increase the depletion of this air mass. A modeling study by Herold and Lohmann (2009) suggests that the continentality was dominant in determining the east-west gradient of the isotopic rainfall composition in Africa during the Eemian period.

The isotopic composition of precipitation over a region is also influenced by changes in the source of moisture due to the changes in atmospheric circulation and the seasonality of precipitation (Lewis et al., 2010; Pausata et al., 2011b). Also the

4.2. Introduction

extent of condensation and fractionation undergone by the air parcel along its ad-vection from the source may increase with the distance to the destination, leading to lower isotope values in precipitation derived from more distantly sourced vapor.

Model and observational studies have analyzed the variability of isotopes in eastern Africa (Levin et al., 2009; Lewis et al., 2010; Tierney et al., 2011a). Observations (Levin et al., 2009) indicate that the comparatively high observed values of hydro-gen isotope composition of precipitation (δDprecip) over eastern equatorial Africa are caused by the enriched moisture originating from the Congo Basin. As such, the CAB serves as an isotopic divide that separates two distinct moisture sources to the region - the Indian Ocean, a relatively depleted moisture source and the recycled, enriched continental moisture (Levin et al., 2009). Likewise, a modeling study by Lewis et al. (2010) concludes that around Lake Tanganyika (eastern Africa) during the HS1, the shift from an Indian Ocean dominated source to a strongly continental and Atlantic-influenced source contributed to the isotope variability through changes in the degree of pre-fractionation and the relative enrichment of the nonfractionat-ing (e.g., durnonfractionat-ing transpiration) continental moisture source. Similarly,Tierney et al.

(2011a) argue that the negativeδD proxy data anomalies from lake Challa in tropical eastern Africa signify the intensity of the East African monsoon circulation rather than the local amount effect.

It has been shown that the compound-specific stable hydrogen isotope compo-sition (δD values) of sedimentary n-alkanes (originating from the epicuticular wax layer of terrestrial plants) can be used for reconstructing past changes in the trop-ical hydrologtrop-ical cycle (Sauer et al., 2001; Schefuß et al., 2005; Collins et al., 2011;

Tierney et al., 2011a). The plant waxes get enriched in deuterium with increased evapo-transpiration, also the soil water gets enriched with the heavier isotope under arid conditions. Thus, the δD values derived from plant waxes correspond to the evaporation-precipitation balance and can be used as a proxy for past changes in